A fast battery charging system applies high charging current to a battery to minimize charging time. The fast charging process is terminated when the battery reaches its full capacity. It is extremely important for the fast charging system to have means for accurately knowing when the battery is fully charged; otherwise the charger will either overcharge the battery or terminate the charging prematurely. The prior art teaches several methods for terminating fast charging of a battery.
One method is to calculate the number of ampere-hours needed to fully charge the battery. The charging time is measured and charging is terminated when the calculated value is reached. The battery must be discharged completely. In addition, the correct ampere-hour value must be obtained. Any mistake or not fully discharging the battery could cause the battery to be overcharged.
Other methods disclosed in the prior art include the use of negative voltage slope termination and delta peak techniques. When NiCd or NiMH batteries are charged with high current, the voltage will increase with time until it reaches full capacity When the battery reaches full capacity, the voltage reaches a peak, levels out, then starts to decline if charging current is still applied. FIG. 4 shows a typical voltage-time curve of a battery under fast charging. Overcharging the battery will produce oxygen on the battery's negative electrode. This oxygen is consumed by the negative electrode and the battery will heat up. The overcharging causes the battery's voltage to drop from Peak point 620 to Overpeak point 640. The voltage drop signals that the battery is already fully charged.
A slope monitoring technique looks at the battery's voltage-time curve and divides it into several sections according to the voltage change rate. The reflection point in the voltage time curve is used as a basis for stopping the charging of the battery. The reflection point used is the point at which the first voltage derivative peaks. The goal is to stop the fast charging while the voltage of the battery is still increasing. There is not, however, a strong relationship between this reflection point and the point when the battery is fully charged. Further, it is hard to find this reflection point accurately due to the precision of measurement required.
Another slope monitoring technique which can charge different voltage rate batteries uses a logarithmic analog-to-digital converter. A voltage divider is used to scale down the high battery voltage to the 5 volt range. The logarithmic analog-to-digital converter is then used to determine the battery's voltage. The logarithmic analog-to-digital converter is actually an RC charging and discharging timing and comparator network and the accuracy of the voltage measurement is poor.
There are a number of variations of the delta peak method. One method is to use a switched power supply and microcontroller in conjunction with an RC charging and discharging circuit and a comparator which measures the battery voltage. The battery voltage could be as high as 13V during charging. A voltage divider is used to scale down the voltage to the 0-5V range. In order to measure the battery voltage, the charging current must be cut off for short periods of time. The charging process is controlled by measuring the charging voltage continuously or at specific time intervals. The battery's voltage range is predetermined. When the delta peak value is greater than the predetermined value, the fast charging is terminated.
Alternatively, the battery's voltage is not predetermined. The previous voltage measurement is used to determine how to select the divider to scale down the battery's voltage measurement into the desired zone. The fast charging is stopped by the time behavior which indicates charging is completed. The time behavior is monitored by the second order time derivative of the battery's voltage. It is well know that the voltage measurement is subject to noise. Any derivation of a noise contaminated signal will be subject to a very serious noise component. This method is complicated and expensive to manufacture.
Another delta peak method for fast charging termination takes advantage of the fact that the voltage versus time curve of a typical battery under charge is not a smooth, substantially linear waveform. As a result, deviations and fluctuations in the battery's voltage will cause the prior known systems to detect false peak voltages and prematurely terminate fast charging. U.S. Pat. No. 5,177,427 issued to Julian J. Bugaj discloses applying a linear time varying causal equation Yn=a*Y+b*Y0, where Yn is the derived battery voltage, "a" and "b" are arbitrary constant coefficients, X is a previously derived battery voltage, and Y0 is the instantaneous battery voltage measurement. This equation is actually just a first order digital filter Yn=a*Yn-1+b*Xn, where Yn is the current derived battery's voltage and Xn is the current voltage measurement. It only gets rid of some of the measurement noise. This charging system still uses a voltage divider to scale down the high battery voltage. The capacitor's charging and discharging time is used to measure the battery's voltage, resulting in a poor voltage measurement. This patent also discloses that after a peak has been detected, a counter will start a count of up to five minutes. If there is any increase in the battery's voltage during this time period the counter will reset. If the counter reaches five minutes, the fast charging will be stopped. Stopping the fast charging five minutes after the first delta peak will lead to excessive overcharging.
U.S. Pat. No. 5,557,190 issued to Daniele C. Brotto discloses the use of the falling voltage slope or delta peak charge termination technique with a noise immunity improvement. Noise is particularly troublesome in the flat regions of the voltage-time curve since a momentary drop in voltage due to noise can be misinterpreted as a falling voltage or a negative peak, resulting in premature termination of the charging. The improvement is to disable the voltage slope charge termination technique during times when the signal to noise ratio is comparatively low. In addition, Brotto attempts to improve the signal to noise ratio by reducing the sampling rate during the flat region of the voltage time curve. Since a voltage divider and logarithmic analog-to-digital converter are still used to measure the battery's voltage, the system suffers the same poor accuracy discussed above.
All of the above prior art allows overcharging of batteries under fast charge. Overcharging of a battery will produce oxygen on the battery's positive electrode. This oxygen is then consumed by the negative electrode. The battery heats up, causing the voltage drop which is used by the above methods to stop the charging. At that point, the battery has already been overcharged. Although battery manufacturers design batteries with extra negative electrode material, eventually the overcharging irreversibly consumes the negative electrode and the battery's capacity decreases.
U.S. Pat. No. 5,694,023, issued to Yury M. Porazhansky, discloses a termination technique which does not rely on the use of the delta peak. A charging pulse is applied, providing an average charging current. A first depolarizing discharging pulse is applied, followed by a first rest period. The voltage is measured at a predetermined point during the first rest period. A second depolarizing discharging pulse is applied, followed by a second rest period. Again, the voltage is measured at a predetermined point during the second rest period. The voltage difference between the first and second voltage is calculated. If the resulting difference is greater than a predetermined amount, fast charging is terminated. There are two problems associated with this method. First, this method slows down the charging process since for every charging pulse there will be two discharging pulses and two rest periods. Secondly, the voltage difference between the first and second voltage does not have a strong correlation to a battery being fully charged; rather, it is closely related to the capacity of the battery being charged and the duration of the second discharge pulse.
The prior arts suffer from a number of voltage measurement problems. First, cheap switched power supplies are used which introduces noise to the circuit. This noise will contaminate the batteries voltage, making the voltage measurement inaccurate. Second, 8-bit analog to digital converters are used which have a resolution of 20 mV with a 5V reference. A voltage divider is used to reduce the battery's voltage to the range of the analog to digital converter measurement. If you charge a 9.6V battery for example, the voltage could reach as high as 13V. The battery's voltage needs to be divided by 3 and then feed to the analog to digital converter. The resulting voltage measurement resolution then is actually 60 mV. Further, noise may cause the charger to stop charging prematurely so that the actual delta threshold used by need to be 120 mV. In order to determine when to terminate the fast charge, the delta peak charger must overcharge the battery in order to get the voltage drop. If the resolution is 120 mV, the battery must be overcharged until a 120 mV voltage drop is obtained. This allows significant overcharging of the battery. Third, there is commonly required an interruption to the charging process in order to perform voltage measurement. Fourth, some poor quality or old batteries will produce a false delta peak, causing charging to be prematurely terminated.